New Mediation Analysis

Although, as described above, there were no direct effects of the
benevolent sexism condition (compared to the control condition)
on attitudes, more recent treatments of mediation suggest this link
might not be necessary and that an indirect effect may nonetheless
exist (e.g., MacKinnon & Fairchild, 2009; Shrout & Bolger, 2002).
In line with this recent literature on mediation, we tested for an
indirect effect and it was significant (indirect effect .14, 95% CI
[.02, .28]). Thus, there was an indirect effect of the benevolent
sexism condition (compared to the control condition) on attitudes
via compassion. Taken together, these results provide evidence
that activating benevolent sexist stereotypes leads to higher experience
of compassion, which in turn leads to more support for a
gender-based EE policy.

Hideg, I., & Ferris, D. L. (2016). The Compassionate Sexist? How Benevolent Sexism Promotes and Undermines Gender Equality in the Workplace.